A green September is framed as a bullish setup for October: green Septembers have been followed by green Octobers more than 60% of the time over the last 20 years, even though the longer 100-year sample is mixed. The tape is still led by AI, memory, and related infra, with NVDA, MU, SK Hynix, and several IA13 names tied to whether earnings, guidance, and AI-fear headlines keep the current uptrend intact. Rates, oil, and yields are the main macro crosscurrents, but the call is still to own AI names plus some crypto, while tracking BTC 82k support, TSLA/SPX pair mechanics, and SpaceX/Tesla valuation spread behavior.
Green Septembers have been a mixed long-run signal: 44 of the last 100 were green and 54 red, but over the last 20 years October has been green more than 60% of the time after a green September.
The cited red-October examples are 2018, 2012, and 2009, with 2009 tied to the global financial crisis, 2012 to a recovery, and 2018 to a weak financial year and 2024 rate hikes at the end of the rate-hike cycle.
10-year yields were cited at 5.285% and rising like oil, U.S. oil was above 90, Brent ran to around 109 before falling back under 100, and a 70% chance of an October rate hike was assigned.
QQQ printed an all-time high on the AI bid, while SPX was also at an all-time high but had a new sell signal and a down-trending chart.
Volatility mean-reverted from 18 to 15 and was described as back to neutral.
BTC support at 82k was called critical, with resistance at 87.3k, a buy signal around 82.6k, and a move framed as manipulated after a roughly $1,000 drop before a top was confirmed.
BTC was expected to hold above 82k, had traded above 84k, reached 85.6k, and then sold off; relative strength versus gold was described as unchanged, with BTC nearly back on a green line watched since the start of the year.
Gold was said to have been hit harder than BTC over the last seven days even while BTC was down.
ETH was chopping sideways with resistance at $2,700.
SOL had beaten ETH by 18% over the last 30 days, was cited around 118–119 before slipping to about 117, had resistance at the same level as last week, support at 112, and a 52-week range high of 237 still far above spot.
Micro Strategy had its prior gap filled on the nose, STRC was near 100 for the first time in months, and the support idea for STRC was tied to 'Sailor' buying it back to 100.
The Tesla/SpaceX pair chart was called critical for merger valuation: the pair was 2.31, described as extremely low, with SpaceX resistance around 155 and a recurring loop where it hits 156–157, falls to 146–145, then mean-reverts higher.
The TSLA/SpaceX pair broke down through a 234–235 box while TSLA was about 350 then 354 in after-hours, and the move was attributed to TSLA moving up.
NVDA had resistance near a green line, no sell flag, and an uptrend that still looked intact; another level cited was 236, with no hedge/buy/sell action in no-man’s-land and an expectation that it keeps moving up for a while.
NVDA entry levels were given at 206 or 208, with the claim that it bounces off the 200-day moving average and should not fall below 200 unless something really bad happens to AI.
SK Hynix was charted on the Korean exchange because the new Nasdaq listing has limited info; the chart has shown an up blue trend since Aug. 18 with no sell signal, and the conditional call is that bad MU earnings would push it to 1.6 million while good earnings keep it moving up.
MU delivered a beat across revenue, EPS, net income, and gross margin: 54.2 vs 50 billion, 33.4 vs 31, 38.48 vs 36, and 87% vs 86%; despite that, the stock only moved up three bucks after the initial reaction.
MU was also tied to a July 29 dip at 738 that was said to be up 46% in 60 days if bought there, while the current guidance was to avoid buying here and wait for 700 or 800.
ALAB was called bullish under 265 earlier, but by Sept. 23 it was at 355 and described as out of juice, with the claim that it would not move up anymore after that run.
AMD had a sell signal at 640, a trend turn at 612, an all-time high at 645, and a bounce around the 600 level from 604 to 600.
GOOGL dipped to 325–326 as a buy zone, tripped a sell flag that took it to 333, then recovered to 340; it was expected to grind higher for the next two or three months and was compared to Buffett-style safety.
PLTR had resistance at 200, hedge coverage at 184, moved up to around 191, folded back, then ran toward 194; the claim is that it heads to 200 and is 200 or above in Q4.
PLTR was also called the weakest and slowest mover of the IA13, with a buy zone at 337, current price 354, and a path back to all-time highs requiring a 41% move.
Tesla was framed as benefiting from expensive oil and high gas/diesel: U.S. showrooms were said to have no cars, Model Y availability was pushed to Q1 2027, European sales were up 3x in some countries, and China was called red hot.
Tesla/FSD was tied to TikTok influencers driving purchases, with FSD said to have broken 15 billion miles and those three influencers representing 99% of miles driven over the last year.
If MU earnings are bad, SK Hynix and the AI memory trade were expected to weaken; if good, the AI hardware complex was expected to keep climbing.
Buying the I13 names 30–40% off all-time highs, plus AI names and a little crypto, was the preferred positioning because high rates, oil, and diesel shortages were seen as being ignored while the economy is driven by AI.
BU was described as 16-year cheap at $87 after a drop into the $85–86 area, with the trend down, a sell signal, slowing growth, lost money, and revenue missing by only about a million dollars and earnings by 1.67%.
Crypto was expected to rise because fiat is going to zero and AI agents will need rails like Solana and USDC to pay, while AI compute, energy, memory, and GPU bottlenecks were expected to intensify from here.